linux-kernel - Re: [PATCH 6/6] cpufreq:amd-pstate: add quirk for the pstate CPPC capabilities missing

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite for Android: free password hash cracker in your pocket

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <49825045-bce1-4772-b7c4-2e116e19cbaf@amd.com>
Date: Wed, 31 Jan 2024 15:59:32 -0600
From: Mario Limonciello <mario.limonciello@....com>
To: Perry Yuan <perry.yuan@....com>, rafael.j.wysocki@...el.com,
 viresh.kumar@...aro.org, Ray.Huang@....com, gautham.shenoy@....com,
 Borislav.Petkov@....com
Cc: Alexander.Deucher@....com, Xinmei.Huang@....com, Xiaojian.Du@....com,
 Li.Meng@....com, linux-pm@...r.kernel.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH 6/6] cpufreq:amd-pstate: add quirk for the pstate CPPC
 capabilities missing

On 1/31/2024 02:50, Perry Yuan wrote:
> Add quirk table to get CPPC capabilities issue fixed by providing
> correct perf or frequency values while driver loading.
> 
> If CPPC capabilities are not defined in the ACPI tables or wrongly
> defined by platform firmware, it needs to use quick to get those

s/quick/quirk/

> issues fixed with correct workaround values to make pstate driver

to allow the pstate driver

> can be loaded even though there are CPPC capabilities errors.
> 
> Signed-off-by: Perry Yuan <perry.yuan@....com>
> ---
>   drivers/cpufreq/amd-pstate.c | 51 +++++++++++++++++++++++++++++++-----
>   include/linux/amd-pstate.h   |  6 +++++
>   2 files changed, 51 insertions(+), 6 deletions(-)
> 
> diff --git a/drivers/cpufreq/amd-pstate.c b/drivers/cpufreq/amd-pstate.c
> index eea2e192d748..cbc415af0f08 100644
> --- a/drivers/cpufreq/amd-pstate.c
> +++ b/drivers/cpufreq/amd-pstate.c
> @@ -64,6 +64,7 @@ static struct cpufreq_driver amd_pstate_driver;
>   static struct cpufreq_driver amd_pstate_epp_driver;
>   static int cppc_state = AMD_PSTATE_UNDEFINED;
>   static bool cppc_enabled;
> +static struct quirk_entry *quirks;
>   
>   /**
>    * struct global_params - Global parameters, mostly tunable via sysfs.
> @@ -121,6 +122,32 @@ static unsigned int epp_values[] = {
>   
>   typedef int (*cppc_mode_transition_fn)(int);
>   
> +static struct quirk_entry quirk_amd_7k62 = {
> +	.nominal_freq = 2600,
> +	.lowest_freq = 550,
> +};
> +
> +static int __init dmi_matched(const struct dmi_system_id *dmi)
> +{
> +	quirks = dmi->driver_data; > +

Under the presumption that the quirk list will grow as more buggy older 
machines are identified, maybe it's worth having a dyndbg or info 
statement here to indicate that it's using quirked values for dmi->ident.

> +	return 1;
> +}
> +
> +static const struct dmi_system_id amd_pstate_quirks_table[] __initconst = {
> +	{
> +		.callback = dmi_matched,
> +		.ident = "AMD EPYC 7K62",
> +		.matches = {
> +			DMI_MATCH(DMI_PRODUCT_VERSION, "C1"),
> +			DMI_MATCH(DMI_PRODUCT_SERIAL, "FX19911000028"),

This is way too specific isn't it?  It would only load on that single 
system I would expect.  But I think you want to have an entry that 
matches the DMI_PRODUCT_NAME or DMI_FAMILY_NAME instead most likely.

> +		},
> +		.driver_data = &quirk_amd_7k62,
> +	},
> +	{}
> +};
> +MODULE_DEVICE_TABLE(dmi, amd_pstate_quirks_table);
> +
>   static inline int get_mode_idx_from_str(const char *str, size_t size)
>   {
>   	int i;
> @@ -581,13 +608,19 @@ static void amd_pstate_adjust_perf(unsigned int cpu,
>   static int amd_get_min_freq(struct amd_cpudata *cpudata)
>   {
>   	struct cppc_perf_caps cppc_perf;
> +	u32 lowest_freq;
>   
>   	int ret = cppc_get_perf_caps(cpudata->cpu, &cppc_perf);
>   	if (ret)
>   		return ret;
>   
> +	if (quirks && quirks->lowest_freq)
> +		lowest_freq = quirks->lowest_freq;
> +	else
> +		lowest_freq = cppc_perf.lowest_freq;
> +
>   	/* Switch to khz */
> -	return cppc_perf.lowest_freq * 1000;
> +	return lowest_freq * 1000;
>   }
>   
>   static int amd_get_max_freq(struct amd_cpudata *cpudata)
> @@ -619,13 +652,14 @@ static int amd_get_max_freq(struct amd_cpudata *cpudata)
>   
>   static int amd_get_nominal_freq(struct amd_cpudata *cpudata)
>   {
> -	struct cppc_perf_caps cppc_perf;
> +	u32 nominal_freq;
>   
> -	int ret = cppc_get_perf_caps(cpudata->cpu, &cppc_perf);
> -	if (ret)
> -		return ret;

I think this hunk should be earlier in your series.  Conceptually it 
seems to match what patch 3 does to me.

> +	if (quirks && quirks->nominal_freq)
> +		nominal_freq = quirks->nominal_freq;
> +	else
> +		nominal_freq = READ_ONCE(cpudata->nominal_freq);
>   
> -	return cppc_perf.nominal_freq;
> +	return nominal_freq;
>   }
>   
>   static int amd_get_lowest_nonlinear_freq(struct amd_cpudata *cpudata)
> @@ -1623,6 +1657,11 @@ static int __init amd_pstate_init(void)
>   	if (cpufreq_get_current_driver())
>   		return -EEXIST;
>   
> +	quirks = NULL;
> +
> +	/* check if this machine need CPPC quirks */
> +	dmi_check_system(amd_pstate_quirks_table);
> +
>   	switch (cppc_state) {
>   	case AMD_PSTATE_UNDEFINED:
>   		/* Disable on the following configs by default:
> diff --git a/include/linux/amd-pstate.h b/include/linux/amd-pstate.h
> index 446394f84606..ea80f961607d 100644
> --- a/include/linux/amd-pstate.h
> +++ b/include/linux/amd-pstate.h
> @@ -110,4 +110,10 @@ static const char * const amd_pstate_mode_string[] = {
>   	[AMD_PSTATE_GUIDED]      = "guided",
>   	NULL,
>   };
> +
> +struct quirk_entry {
> +	u32 nominal_freq;
> +	u32 lowest_freq;
> +};
> +
>   #endif /* _LINUX_AMD_PSTATE_H */